`migrate()` throws error when `time` values have overlapping characters

[mthomas-ketchbrook]

Problem

When one of the distinct values for the time argument contains the other distinct value, the column name replacement logic will cause migrate() to throw an error stating that it "Can't subset elements that don't exist". For example, if one of the time values is "M1" and the other is "M12", the first set of gsub() logic will affect the column names for both of the distinct values for time.

Reproducible Example

library(migrate)

mock_credit |>
  dplyr::mutate(
    problematic_time = dplyr::case_when(
      date == as.Date("2020-06-30") ~ "time_1",
      date == as.Date("2020-09-30") ~ "time_12"   # this string contains "time_1"
    )
  ) |>
  migrate(
    id = customer_id,
    time = problematic_time,
    state = risk_rating,
  )

throws the following error:

ℹ Migrating from time_1 to time_12
Error in `dplyr::group_by()`:
ℹ In argument: `dplyr::across(dplyr::all_of(c(state_start_name, state_end_name)))`.
Caused by error in `across()`:
ℹ In argument: `dplyr::all_of(c(state_start_name, state_end_name))`.
Caused by error in `dplyr::all_of()`:
! Can't subset elements that don't exist.
✖ Element `risk_rating_end` doesn't exist.
Run `rlang::last_trace()` to see where the error occurred.

Possible Solution

We need to be stricter about how we name/rename these columns. Perhaps we should create these column names without performing any string replacement.

[mthomas-ketchbrook]

Closed by #22